A method and system for non-contact passability estimation of a Mars rover based on thermal inertia

Abstract

The application belongs to the technical field of data processing, and discloses a Mars rover non-contact passability prediction method and system based on thermal inertia, which comprises the following steps: obtaining the ground temperature time sequence and environmental parameters of a target area; constructing a thermal inertia estimation model, inputting the ground temperature time sequence and environmental parameters into the thermal inertia estimation model, and obtaining the thermal inertia estimation value of the soil; constructing a thermal inertia and passability index relationship model, inputting the thermal inertia estimation value of the soil into the thermal inertia and passability index relationship model, and outputting a predicted passability index; determining the passability state level according to the predicted passability index, and generating a navigation control suggestion, so that the Mars rover passes the target area according to the navigation control suggestion of the Mars rover. The application can effectively improve the evaluation accuracy of the Mars rover passing the target area by extracting the thermal inertia estimation value of the soil of the target area to predict the passability index of the Mars rover, and ensures the driving safety of the Mars rover.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a non-contact method, system, terminal, and computer-readable storage medium for predicting the passability of a Mars rover based on thermal inertia. Background Technology

[0002] Mars is largely covered by desert, and its atmosphere creates a variety of complex and dangerous terrains, including undulating hills, soft sand, and a hard, thin crust. These terrains are visually deceptive; for example, a hard, thin crust measuring millimeters to centimeters may lie tens of centimeters deep beneath it; a seemingly flat and homogeneous sandy area may exhibit orders of magnitude differences in density and load-bearing capacity in different regions. This inconsistency between surface morphology and underlying mechanical properties makes Mars rovers highly susceptible to sinking or severely slipping in seemingly safe areas, posing a continuous and significant threat to their driving safety.

[0003] To address the aforementioned issues, current Mars rover mobility predictions primarily rely on two technical approaches: the first is geometric mobility assessment based on visual sensors such as visible light cameras and lidar; the second is dynamic data inversion based on the contact between the wheels and the soil.

[0004] However, neither of these two methods can perceive or predict the internal mechanical properties of the soil in advance, resulting in low accuracy of the rover's passability assessment and affecting the rover's driving safety.

[0005] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0006] The main objective of this invention is to provide a non-contact method and system for predicting the passability of Mars rovers based on thermal inertia. This aims to solve the problem that existing technologies cannot perceive and predict the internal mechanical properties of Martian soil, resulting in low accuracy in assessing the passability of Mars rovers and affecting their driving safety.

[0007] To achieve the above objectives, the present invention provides a non-contact Mars rover passability prediction method based on thermal inertia, the method comprising the following steps:

[0008] Identify the target area and obtain the time series of surface temperature and environmental parameters of the target area;

[0009] A thermal inertia estimation model was constructed, and the surface temperature time series and the environmental parameters were input into the thermal inertia estimation model to obtain the estimated value of the thermal inertia of the fire-soil mixture.

[0010] A model relating thermal inertia to passability index is constructed. The estimated thermal inertia of the fire soil is input into the model relating thermal inertia to passability index, and the predicted passability index is output.

[0011] The passability status level is determined based on the predicted passability index, and navigation control suggestions are generated based on the passability status level, so that the rover can pass through the target area according to the rover navigation control suggestions.

[0012] Optionally, the non-contact mobility prediction method for Mars rovers based on thermal inertia, wherein determining the target area and obtaining the time series of surface temperature and environmental parameters of the target area specifically includes:

[0013] Determine the target region, obtain the region shape of the target region, and determine the time resolution based on the region shape;

[0014] A preset illumination mode is determined, and the thermal infrared imaging equipment on the Mars rover is used to detect thermal radiation in the target area according to the time resolution and the preset illumination mode to obtain multiple frames of thermal infrared images of the target area within a preset time period.

[0015] The preset lighting modes include natural lighting mode and active thermal radiation mode;

[0016] Extract the time sequence of surface temperature variation from the multi-frame thermal infrared images to obtain the surface temperature time series;

[0017] Obtain the environmental parameters of the target area.

[0018] Optionally, the non-contact passability prediction method for Mars rovers based on thermal inertia, wherein extracting the time sequence of surface temperature changes in the multi-frame thermal infrared images to obtain the surface temperature time series specifically includes:

[0019] The multi-frame thermal infrared images are preprocessed, including fixed defect detection processing, noise reduction processing, alignment processing, and feature point extraction processing, to obtain a preprocessed image.

[0020] Calculate the mean and standard deviation of the temperature of the effective pixels in each frame of the preprocessed image, and obtain the target temperature value based on the mean and standard deviation of the temperature;

[0021] The target temperature values ​​are arranged in chronological order to obtain the initial temperature time series;

[0022] The initial temperature time series is smoothed to obtain the surface temperature time series.

[0023] Optionally, in the aforementioned non-contact passability prediction method for Mars rovers based on thermal inertia, the environmental parameters include albedo.

[0024] The acquisition of environmental parameters of the target area specifically includes:

[0025] The target area is imaged and radiometrically calibrated using a multispectral camera on the Mars rover to obtain an albedo image. The average albedo in the albedo image is then calculated to obtain the albedo.

[0026] Alternatively, a high-resolution visible light camera or color camera carried by a Mars rover can be used to image and perform radiometric calibration on the terrain areas with shadow boundaries in the target area to obtain a reflectance image;

[0027] The reflectance image is divided into a direct illumination area and a shadow area using an image processing algorithm, and the average reflectance of the direct illumination area and the shadow area is calculated.

[0028] The average reflectance of the shaded area is corrected for atmospheric scattering to obtain the albedo.

[0029] Alternatively, a pre-trained deep convolutional neural network can be constructed to obtain a visible light image of the target region, and the visible light image can be analyzed using the pre-trained deep convolutional neural network to obtain the albedo.

[0030] Optionally, the non-contact passability prediction method for Mars rovers based on thermal inertia, wherein constructing a thermal inertia estimation model and inputting the surface temperature time series and the environmental parameters into the thermal inertia estimation model to obtain the estimated value of the Mars soil thermal inertia, specifically includes:

[0031] A thermal inertia estimation model is constructed based on the Martian surface heat balance equation and heat conduction equation. The surface temperature time series and the environmental parameters are input into the thermal inertia estimation model, and the soil thermal inertia is calculated based on the surface temperature time series and the environmental parameters through the thermal inertia estimation model to obtain the estimated value of Martian soil thermal inertia.

[0032] The expression for the thermal inertia estimation model is as follows:

[0033] ;

[0034] in, This is an estimated value for the thermal inertia of volcanic soil. This is a comprehensive correction factor related to the environment of the target area. The albedo is one of the environmental parameters. The maximum surface temperature difference within a selected time period in the surface temperature time series.

[0035] Optionally, the non-contact passability prediction method for Mars rovers based on thermal inertia, wherein constructing a relationship model between thermal inertia and passability indicators, inputting the estimated thermal inertia of the Mars soil into the relationship model, and outputting predicted passability indicators, specifically includes:

[0036] Multiple sets of simulated fire soil samples were prepared, and the measured values ​​of thermal inertia and mechanical parameters of the simulated fire soil samples were obtained.

[0037] A dataset is constructed based on the measured values ​​of thermal inertia and mechanical parameters, and the dataset is divided into a training set, a validation set, and a test set.

[0038] Determine a multinomial regression model, and train the multinomial regression model using the training set to obtain an initial model;

[0039] Alternatively, a support vector regression model or a gradient boosting regression tree model can be used. The support vector regression model or the gradient boosting regression tree model can be trained based on the training set to obtain an initial model.

[0040] The initial model is cross-validated based on the validation set, and the model is optimized based on the test set to obtain a model relating thermal inertia and passability index.

[0041] The estimated thermal inertia of the fire and soil is input into the relationship model between thermal inertia and passability index, and the predicted passability index is output.

[0042] Optionally, the non-contact passability prediction method for Mars rovers based on thermal inertia, wherein determining the passability status level based on the predicted passability index and generating navigation control suggestions based on the passability status level, enabling the Mars rovers to pass through the target area according to the Mars rovers' navigation control suggestions, specifically includes:

[0043] Determine the preset safety margin and preset safety threshold;

[0044] If the predicted passability index is greater than or equal to the sum of the preset safety margin and the preset safety threshold, then the target area is determined to be at the first passability status level.

[0045] If the predicted passability index is greater than the preset safety threshold and less than the sum of the preset safety margin and the preset safety threshold, then the target area is determined to be at the second passability status level.

[0046] If the predicted passability index is less than the preset safety threshold, the target area is determined to be at the third passability status level.

[0047] Based on the first passability status level, the second passability status level, and the third passability status level, corresponding navigation control suggestions are generated respectively, enabling the Mars rover to pass through the target area according to the Mars rover navigation control suggestions.

[0048] Furthermore, to achieve the above objectives, the present invention also provides a non-contact Mars rover passability prediction system based on thermal inertia, wherein the non-contact Mars rover passability prediction system based on thermal inertia includes:

[0049] The regional data acquisition module is used to determine the target area and acquire the land surface temperature time series and environmental parameters of the target area;

[0050] The thermal inertia estimation module is used to construct a thermal inertia estimation model and input the surface temperature time series and the environmental parameters into the thermal inertia estimation model to obtain the thermal inertia estimation value of the fire soil;

[0051] The passability index output module is used to construct a model relating thermal inertia and passability index. The estimated value of the fire soil thermal inertia is input into the model relating thermal inertia and passability index, and the predicted passability index is output.

[0052] The navigation suggestion generation module is used to determine the passability status level based on the predicted passability index, and generate navigation control suggestions based on the passability status level, so that the rover can pass through the target area according to the rover navigation control suggestions.

[0053] Furthermore, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a Mars rover non-contact passability prediction program stored in the memory and executable on the processor, wherein when the Mars rover non-contact passability prediction program based on thermal inertia is executed by the processor, it implements the steps of the Mars rover non-contact passability prediction method based on thermal inertia as described above.

[0054] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a Mars rover non-contact passability prediction program based on thermal inertia, and when the Mars rover non-contact passability prediction program based on thermal inertia is executed by a processor, it implements the steps of the Mars rover non-contact passability prediction method based on thermal inertia as described above.

[0055] In this invention, a target area is determined, and the surface temperature time series and environmental parameters of the target area are obtained. A thermal inertia estimation model is constructed, and the surface temperature time series and environmental parameters are input into the thermal inertia estimation model to obtain the estimated value of the fire-soil thermal inertia. A relationship model between thermal inertia and passability index is constructed, and the estimated value of the fire-soil thermal inertia is input into the thermal inertia-passability index relationship model to output the predicted passability index. The passability status level is determined according to the predicted passability index, and navigation control suggestions are generated according to the passability status level, enabling the Mars rover to pass through the target area according to the Mars rover navigation control suggestions. This invention, by constructing a thermal inertia estimation model, can extract the estimated value of the thermal inertia of the target area and predict the passability index of the Mars rover based on the estimated value of the thermal inertia, which can effectively improve the assessment accuracy of the Mars rover passing through the target area and ensure the driving safety of the Mars rover. Attached Figure Description

[0056] Figure 1 This is a flowchart of a preferred embodiment of the Mars rover non-contact passability prediction method based on thermal inertia of the present invention;

[0057] Figure 2 This is a structural diagram of a preferred embodiment of the Mars rover non-contact passability prediction system based on thermal inertia of the present invention;

[0058] Figure 3 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0060] The 21st century marks a new era for humankind's comprehensive space exploration and development, with Mars being one of the key planets targeted for exploration. Mars is largely covered by desert, and its atmosphere creates a variety of complex and dangerous terrains, including undulating hills, soft sand, and hard, thin crusts. These terrains are visually deceptive; for example, a hard, thin crust measuring millimeters to centimeters may conceal tens of centimeters of loose sand. A seemingly flat and homogeneous sandy area may exhibit order-of-magnitude differences in density and load-bearing capacity across different regions. This inconsistency between surface morphology and underlying mechanical properties makes Mars rovers highly susceptible to sinking or severe slippage in seemingly safe areas, posing a continuous and significant threat to their operational safety. Historically, a certain Mars rover has encountered such dangers multiple times, leading to lengthy mission delays and even permanent loss of mobility, resulting in incalculable scientific losses.

[0061] To ensure the safe operation of Mars rovers, current rovers rely primarily on two technical approaches for assessing their traversability. The first approach relies on geometric traversability assessment using visual sensors such as visible light cameras and lidar. This method uses 3D reconstruction to obtain information on terrain elevation, slope, and roughness, identifying obvious geometric obstacles such as steep slopes, large rocks, and deep pits. However, this method only perceives surface topography and is completely incapable of sensing the internal mechanical properties of the soil (such as compressive strength, shear strength, and cohesion), which are crucial for traversability. Therefore, it is ineffective against soft sandy soils and thin-crust terrains—topography that is "soft in appearance but treacherous in reality." The second approach is based on inverse dynamic data analysis after the wheels come into contact with the soil. This method analyzes real-time data such as wheel slip rate, subsidence depth, and drive motor torque or current during wheel movement, combined with a wheel-soil mechanics model, to inversely calculate the mechanical parameters of the soil beneath the wheels. Although this method can provide mechanical information about the contact area, it is essentially a "post-judgment" or "touch-and-go diagnosis." It can only issue an alarm after the rover has entered and may have begun to sink into a certain area. It is completely unable to provide forward warnings for unknown areas that have not yet been touched, thus losing the initiative to avoid risks.

[0062] Thermal inertia Thermal conductivity is a comprehensive parameter characterizing the thermophysical properties of a material, and it is defined as the thermal conductivity of the material. ,density With specific heat capacity square root of the product For granular media such as Martian soil, thermal inertia is closely related to physical properties such as particle size distribution, bulk density (porosity), interparticle contact thermal resistance, and the presence of cementing substances (affecting cohesion). It is noteworthy that these physical properties affecting thermal inertia (density, particle size, cohesion) are precisely the key intrinsic factors determining the macroscopic mechanical behavior of soil, namely its compressive strength (resistance to subsidence) and shear strength (resistance to slippage). This reveals a deep and fundamental correlation between thermal inertia and soil support permeability from a physical mechanism perspective.

[0063] In existing technologies, thermal inertia has been widely used in planetary science, but its application scenarios have inherent limitations. Specifically, it is mainly used as an orbital remote sensing inversion tool. Through thermal infrared spectrometers mounted on orbiters, it acquires large-scale, long-period thermal radiation data of the planetary surface, and then inverts the regional average thermal inertia at a scale of several kilometers to tens of kilometers. This data has been successfully used to create global geological maps, distinguish macroscopic units such as rocks, dunes, and dust, and study climate evolution. However, the spatial resolution of orbital remote sensing thermal inertia data is extremely low (typically hundreds of meters to kilometers), and it reflects spatially averaged thermal characteristics at a relatively deep depth (up to tens of centimeters) below the surface. Its data update cycle is long, making it unable to capture local and transient changes. Therefore, existing orbital thermal inertia products cannot meet the urgent need of Mars rovers to conduct real-time, detailed, and high-precision passability assessments of the terrain within a range of several meters to tens of meters around the rovers and along a specific travel path.

[0064] In summary, how to use Mars rover onboard sensors to quickly and non-contactly estimate the thermal inertia of uncontacted local areas and establish its quantitative relationship with key passability indicators to achieve forward-looking early warning is a technical problem that urgently needs to be solved.

[0065] To address the aforementioned problems, this invention provides a non-contact method and system for predicting the passability of a Mars rover based on thermal inertia, belonging to the field of planetary exploration technology. The method includes: acquiring a time series of surface temperatures in the uncontacted target area ahead of the Mars rover; calculating the regional soil thermal inertia value based on this series using a thermal inertia estimation model; and inputting the thermal inertia value into a pre-generated thermal inertia-passability index relationship model to predict the region's support passability status. The thermal inertia-passability index relationship model is established through ground simulation experiments, characterizing the quantitative relationship between soil thermal inertia and mechanical parameters such as the conic index. The system of this invention includes a thermal sensing module and a data processing and control module. This invention achieves a forward-looking and quantitative assessment of the support passability of soft Martian terrain in a non-contact state, and is particularly adept at identifying visually invisible thin shells, soil coverings, and other dangerous terrain, significantly improving the safety and exploration efficiency of the Mars rover's autonomous navigation.

[0066] Understandably, in Mars exploration, thin-shell terrain (i.e., a hard, thin shell on the order of millimeters to centimeters covered by loose sand and soil tens of centimeters deep) is extremely visually deceptive and poses a significant threat that could cause rovers to sink or slip.

[0067] The identification principle of this invention includes: utilizing differences in thermal response to detect "deep characteristics." Thermal inertia is a comprehensive parameter characterizing the square root of the product of a material's thermal conductivity, density, and specific heat capacity. Soils with different structures (such as hard crusts and loose sand) exhibit completely different surface temperature changes when subjected to thermal excitation. Visual sensors can only perceive the surface topography and cannot penetrate the thin crust to perceive the material beneath; however, thermal inertia can indirectly perceive the density gradient and cohesion within the soil by monitoring the conduction and dissipation processes after the surface is heated.

[0068] The core means of this invention include: active thermal induction mode. In order to accurately identify thin shells, the system will activate an active heat source (such as a focusing lens or an electric heating radiation device) to perform short-term, high-intensity local heating on the suspicious area: the surface is irradiated using the mid- to far-infrared band (5 to 15 μm), which has a higher detection matching degree for shallow materials, and the heating process during irradiation and the cooling process after irradiation are recorded with high temporal resolution (once every 5 seconds).

[0069] The specific identification logic of this invention includes: judgment based on the transient characteristics of heating and cooling. For thin-shell terrain features: if the surface temperature of the target area rises extremely rapidly during heating, but the temperature decays (cools) abnormally slowly after heating stops, it can be identified as a thin-shell terrain. Principle: A rapid heating rate indicates that the surface material is thin and the thermal conductivity below is low. Loose sand hinders the rapid transfer of heat to deeper layers, and heat accumulates on the thin-shell surface. Slow cooling indicates that the loose sand below acts as thermal insulation, and the efficiency of heat diffusion to the interior through conduction is extremely low. If it is a homogeneous, hard terrain, heat will be rapidly transferred to the interior, resulting in a slow temperature rise. This is because hard terrain has good thermal conductivity, and after absorbing heat, it quickly transfers the heat to deeper layers, preventing extreme high temperatures on the surface. After heating stops, heat continues to diffuse to the interior, and the surface temperature decreases slowly. If it is a homogeneous, loose sandy area, due to its extremely low thermal inertia, the surface thermal response is extremely sensitive. The surface temperature rises sharply during heating, and the cooling rate is also relatively rapid during cooling.

[0070] Therefore, once the system detects this abnormal curve of "rapid rise and slow fall," it will immediately execute the following logic, through the mapping model. CI =F( I i This thermal response characteristic is converted into a low load-bearing capacity index and classified as a Level 3 command. This logic effectively solves the problem that visual sensors cannot see through "surface traps." The most direct reference for defining "fast" and "slow" is a database of simulated volcanic soil samples established in advance on the ground. The curves measured by the Mars rover are compared with these "standard curves." For example, if the measured heating slope is close to that of the "loose fine sand" standard sample and the cooling slope is significantly lower than that sample, it is defined as a "thin-shell terrain" feature.

[0071] The preferred embodiment of the present invention describes a non-contact Mars rover passability prediction method based on thermal inertia, such as... Figure 1 As shown, the non-contact Mars rover passability prediction method based on thermal inertia includes the following steps:

[0072] Step S10: Determine the target area and obtain the surface temperature time series and environmental parameters of the target area.

[0073] To address the shortcomings of existing technologies, the present invention aims to provide a non-contact Mars rover passability prediction method and system based on thermal inertia. This method can acquire thermal data of the target terrain using onboard thermal infrared equipment before the Mars rover contacts the terrain, estimate its thermal inertia, and predict its passability indicators based on a pre-established physical relationship model. This provides real-time and reliable decision-making basis for the Mars rover's autonomous path planning and safe navigation.

[0074] Meanwhile, this invention provides a non-contact passability prediction system for implementing the above-mentioned method, integrated on a Mars rover, including a thermal infrared sensing module, an environmental perception module, and a data processing and control module. The thermal infrared sensing module is used to collect thermal infrared image data of the non-contact area; the environmental perception module is used to acquire or estimate the environmental parameters of the current Martian surface; the data processing and control module includes a storage unit, a computing unit, and a decision-making unit, wherein the storage unit is used to store the thermal inertia estimation model and the thermal inertia-passability index relationship model; the computing unit is used to call the model and perform calculations for thermal inertia estimation and passability index prediction; the decision-making unit is used to generate passability judgments and navigation suggestions based on the prediction results. In addition, the system of this invention may also include an active thermal induction module for controllable active heating of the target area.

[0075] The present invention first requires the acquisition and processing of thermal data. The specific implementation steps are as follows: 1. Control the thermal infrared imaging equipment on the Mars rover to continuously monitor the target area that has not yet been contacted, and acquire multiple frames of thermal infrared images of the target area within a selected time period; 2. Extract the time sequence of the surface temperature of the target area from the thermal infrared images; at the same time, acquire or estimate the environmental parameters of the target area. The environmental parameters include at least albedo (preferred in the present invention). In addition, the environmental parameters may also include direct solar irradiance, solar altitude angle and azimuth angle, atmospheric temperature, atmospheric pressure and atmospheric transmittance.

[0076] Specifically, the process involves: determining a target region, acquiring its shape, and determining a temporal resolution based on the shape; determining a preset illumination mode, and using a thermal infrared imaging device on the Mars rover to perform thermal radiation detection on the target region according to the temporal resolution and the preset illumination mode, resulting in multiple frames of thermal infrared images of the target region within a preset time period; wherein the preset illumination mode includes a natural illumination mode and an active thermal radiation mode; preprocessing the multiple frames of thermal infrared images, including fixed defect detection, noise reduction, alignment, and feature point extraction, to obtain a preprocessed image; calculating the mean and standard deviation of the temperature of effective pixels in each frame of the preprocessed image, and obtaining a target temperature value based on the mean and standard deviation; arranging the target temperature values ​​in chronological order to obtain an initial temperature time series; and smoothing the initial temperature time series to obtain a surface temperature time series.

[0077] The process of extracting temperature from thermal infrared images is as follows: Based on the rover's predetermined path, an area within 5 to 10 meters ahead is automatically selected. This area may be rectangular, circular, or polygonal. Different time resolutions are selected according to different modes: natural light is selected every minute to cover the entire selected time period; active thermal radiation mode is selected every 5 seconds to fully cover the heating and cooling process. Each image records the shooting time, camera position, and attitude information. The original images are stored sequentially by number, and corresponding metadata files are also saved. Next, the images are preprocessed to identify fixed bad pixels (pixels that do not affect temperature changes). Random noise in the images is removed using median filtering within a 3×3 or 5×5 pixel window, preserving image edge information while smoothing noise. Since the rover may move slightly, images taken at different times need to be aligned. The feature point method is used to find obvious feature points in each image. The final step is temperature extraction and time series construction. For the target area in each frame, the mean temperature of all effective pixels in the area is calculated, along with the standard deviation of the temperature. Each frame is further processed and arranged sequentially to form a time and temperature sequence. In the quality control and post-processing stage, temperature values ​​that deviate significantly from the normal range are identified, and the time series is smoothed to remove random fluctuations. Finally, a temperature time series file in CSV format (i.e., the surface temperature time series in this invention) is output, which includes data such as time, temperature, and standard deviation.

[0078] Furthermore, for thermal data acquisition, passive natural lighting mode or active thermal induction mode can be selectively adopted according to the real-time needs of the task and environmental conditions. The specific implementation steps are as follows:

[0079] 1. Passive Natural Illumination Mode: This mode utilizes natural Martian solar radiation as the thermal excitation source. After the rover selects the target area to be evaluated (considering resolution, the coverage range is 5 to 10 meters, and the rover needs 1 to 3 meters to come to a complete stop from its cruising speed, so the target area must be at least 3 meters away), the rover controls the thermal infrared camera to continuously monitor the area, recording the natural changes in surface temperature over a complete Martian day cycle or a specific time period (such as from dawn to noon, or from sunset to midnight). The core advantage of this mode is that it is completely passive, requiring no additional energy to intervene in the target area, relying solely on the environmental thermal cycle. However, its data acquisition time is relatively long (several hours to a Martian day), and the evaluation effect is limited by the natural illumination conditions at the time (such as whether it is during a dust storm, whether the solar altitude angle is too low, etc.), making it more suitable for scenarios with ample mission time or where the rover's energy is strictly confidential.

[0080] 2. Active Thermal Induction Mode: To overcome the shortcomings of passive modes, such as long irradiation time and susceptibility to natural conditions, this invention further proposes an active thermal induction mode. In this mode, a dedicated active heat source carried by the Mars rover is used to subject the target area to short-term, high-intensity, and locally controlled irradiation, artificially creating a significant temperature disturbance. A thermal infrared camera is then used to monitor the transient temperature response of the area during irradiation and the cooling process after irradiation is stopped at high frequency.

[0081] Specifically, the selection of the heat source, i.e., the selection of the active heat source, can be as follows:

[0082] a. Concentrating lens system: Composed of a pointing mechanism and an optical lens group, it is used to focus natural sunlight on Mars onto a point or a small area of ​​the target region to achieve instantaneous high-energy-density photothermal conversion. The system is powered by the sun, but the rover achieves active spatial and temporal control of solar energy.

[0083] b. Electric heating radiation device: For example, a radiation source based on resistance wire or semiconductor heating element, powered by the Mars rover's battery, can emit electromagnetic waves in specific bands (such as mid- and far-infrared) to heat the soil. This method is not limited by day or night, and the heating power and time can be precisely controlled.

[0084] Core Advantages and Applicability: The main advantages of the active mode lie in its speed, controllability, and independence from the natural day-night cycle. It can complete "thermal diagnostics" of suspicious locations within minutes at any time the rover is paused. This method stimulates the thermal response of the top and shallow (centimeter-level) soil layers, which is more closely matched to the critical depth of wheel interaction, offering unique advantages for detecting dangerous structures such as thin shells and shallow, loose overburden. This mode is particularly suitable for key decision points in path planning or for rapid verification of visually highly suspicious areas.

[0085] Furthermore, this invention establishes a mode selection and fusion strategy: in practical tasks, the two modes can constitute a complementary cooperative observation strategy.

[0086] a. Routine cruise screening: During long-distance cruises, the passive mode can be mainly relied upon to perform a large-scale, low-frequency scan of the thermal inertia background value of a wide area ahead during driving breaks or nighttime parking, and to establish a regional "thermal characteristic base map".

[0087] b. Detailed investigation and risk assessment of key targets: When passive screening discovers areas with abnormal thermal characteristics, or visual navigation identifies high-risk terrain (such as suspected sand pits or ripples), the active mode is activated to conduct rapid and detailed fixed-point detection to obtain high-temporal-resolution transient thermal response data for accurate passability determination.

[0088] c. Data Fusion and Validation: When conditions permit, two modes can be used sequentially to probe the same area. The full-day cycle data acquired by the passive mode can be used to calibrate and validate the reliability of the thermal inertia parameters retrieved by the active mode under short-term excitation, thereby continuously optimizing the excitation and inversion algorithms of the active mode.

[0089] The target area is imaged and radiometrically calibrated using a multispectral camera mounted on the Mars rover to obtain an albedo image. The average albedo in the albedo image is then calculated to obtain the albedo.

[0090] Furthermore, regarding the acquisition of albedo from environmental parameters, albedo... As one of the key environmental parameters, it can be obtained in one or more of the following ways:

[0091] 1. Real-time inversion based on onboard multispectral or visible light cameras: The multispectral camera onboard the Mars rover can perform multi-band imaging of the target area. By analyzing the reflectance spectrum in the visible to near-infrared bands, the weighted average albedo of the area within the solar spectrum is calculated, which is the albedo. The specific implementation of active thermal radiation is as follows: Before activating the active heat source, the Mars rover first uses its onboard multispectral camera or visible light camera to capture one or more frames of images of the target area. After radiometric calibration, the image data is converted into an albedo image, and the average albedo of the area is calculated as the albedo. .

[0092] Alternatively, a high-resolution visible light camera or color camera mounted on a Mars rover can be used to image and radiometrically calibrate the terrain areas with shadow boundaries in the target region to obtain a reflectance image. An image processing algorithm is then used to divide the reflectance image into directly illuminated areas and shadowed areas, and the average reflectance of the directly illuminated areas and the shadowed areas is calculated. Atmospheric scattering correction is then applied to the average reflectance of the shadowed areas to obtain the albedo.

[0093] 2. Albedo Calculation Based on Reflectance Comparison between Shaded and Illuminated Areas: This method is based on the light scattering theory of Lambertian bodies. It utilizes the difference in reflectance between areas directly illuminated by the sun and areas in shadow within the same scene, combined with geometric parameters such as the angle of solar incidence and the observation angle, to invert the optical scattering characteristics of the Earth's surface, thereby calculating the hemispherical albedo. Its physical basis lies in the fact that reflected light in shadowed areas mainly comes from sky-scattered light and multiple scattering from the surrounding terrain, while illuminated areas include contributions from both direct and scattered light. Specifically, a high-resolution visible light or color camera onboard the Mars rover is used to image terrain areas containing distinct shadow boundaries. The images undergo radiometric calibration to convert them into reflectance images. Then, image processing algorithms are used to divide the images into directly illuminated areas (direct sunlight) and shadowed areas (only illuminated by sky-scattered light and reflected light from the surrounding environment). The reflectance is then extracted and corrected, and the average reflectance of the illuminated and shadowed areas is calculated separately. and Based on the solar altitude angle and observation angle, atmospheric scattering correction is applied to the reflectance of the shaded area to obtain the reflectance component contributed solely by surface scattering. A simplified physical relationship model is then established, expressed as follows:

[0094] ;

[0095] in: Albedo; For the observation angle, Angle of incidence Is with , The relevant correction factors can be determined in advance through ground calibration tests.

[0096] Alternatively, a pre-trained deep convolutional neural network can be constructed to obtain a visible light image of the target region, and the visible light image can be analyzed using the pre-trained deep convolutional neural network to obtain the albedo.

[0097] 3. Albedo Estimation Based on Machine Learning and Image Features: This method utilizes a large amount of Martian surface albedo data (from pre-Mars rover measurements using orbital remote sensing) to train a deep convolutional neural network, enabling it to directly predict surface albedo from single or multiple visible light images. The model learns the complex mapping relationship between visual features such as texture, color, shadows, and terrain undulations in the image and albedo, achieving rapid end-to-end estimation. Specifically, it first collects Martian orbital albedo data (e.g., THEMIS albedo map) and corresponding high-resolution visible light images (e.g., HiRISE images). Then, it performs image registration, cropping, and normalization to form "input image and ground truth albedo" sample pairs. Simultaneously, it can further incorporate spectral and image synthesis data of simulated Martian soil to enhance data diversity. During model training, a convolutional neural network is used, with RGB or multiple image patches as input and the average albedo of the corresponding region as input. The loss function can be mean squared error. Augmentation data can be introduced during training to improve generalization ability. Finally, the trained model is uploaded to the Mars rover's data processing unit. During the rover's journey, visible light images of the area ahead are collected in real time, input into the model, and the albedo prediction value is directly output. At the same time, the uncertainty module can be combined to provide prediction confidence.

[0098] Step S20: Construct a thermal inertia estimation model, and input the surface temperature time series and the environmental parameters into the thermal inertia estimation model to obtain the estimated value of the fire-soil thermal inertia.

[0099] This invention calculates the estimated value of the volcanic thermal inertia of the target area by inputting the surface temperature sequence and the environmental parameters into a thermal inertia estimation model. I The thermal inertia estimation model is constructed based on the Martian surface heat balance equation and the heat conduction equation.

[0100] Specifically, a thermal inertia estimation model is constructed based on the Martian surface heat balance equation and heat conduction equation. The surface temperature time series and the environmental parameters are input into the thermal inertia estimation model, and the soil thermal inertia is calculated based on the surface temperature time series and the environmental parameters through the thermal inertia estimation model to obtain the estimated value of Martian soil thermal inertia.

[0101] The expression for the thermal inertia estimation model is as follows:

[0102] ;

[0103] in, This is an estimated value for the thermal inertia of volcanic soil. This is a comprehensive correction factor related to the environment of the target area. The albedo is one of the environmental parameters. The maximum surface temperature difference within a selected time period in the surface temperature time series (i.e., the difference between the highest and lowest temperatures in the selected surface temperature series).

[0104] Wherein, the correction coefficient It is not a fixed constant, but is based on Martian atmospheric dust and opacity ( ), angle of solar incidence ( ) and Martian season ( The real-time dynamic function of ), where, The expression is:

[0105] ;

[0106] Reason: Atmospheric compensation, introducing transmittance correction based on Beer-Lambert's law, and obtaining current atmospheric dust data through onboard environmental monitoring. The value is used to correct the actual effective irradiance reached by the focusing lens on the earth's surface, eliminating the deviation in the temperature rise curve caused by suspended sand and dust.

[0107] The reason for using this formula in the thermal inertia estimation model on Mars is that... This formula involves many variables, but the Martian environment lacks the conditions and instruments to measure these variables. Therefore, [the following option is chosen]. As a simplified thermal inertia model (i.e., the thermal inertia estimation model in this invention).

[0108] The feasibility of the thermal inertia estimation model in this invention is as follows: Compared to Earth's environment, Mars has a thin, dry, and homogeneous atmosphere. The thin Martian atmosphere has weak energy absorption and scattering effects, allowing solar radiation to reach the surface more directly, and infrared radiation from the surface to dissipate more directly into space. The atmosphere provides minimal resistance to radiative heat transfer between surface particles. Mars lacks the dramatic water cycle (such as heavy rainfall and typhoons), complex weather systems, and extensive cloud cover found on Earth, making its atmospheric thermal state relatively more stable and predictable. Water is a powerful heat carrier, and its phase change and flow greatly complicate heat exchange processes. In the thermal model, there is no need to consider the dramatic spatiotemporal variations in "soil moisture" and its impact on heat capacity and thermal conductivity, as is the case on Earth, nor is it necessary to consider the latent heat effects of evaporative cooling or condensation heating.

[0109] Step S30: Construct a model relating thermal inertia and passability index. Input the estimated thermal inertia of the fire soil into the model relating thermal inertia and passability index, and output the predicted passability index.

[0110] For predicting the passability index, the present invention inputs the estimated thermal inertia value I into a pre-generated thermal inertia-passability index relationship model, thereby outputting the predicted passability index of the target area; wherein, the thermal inertia-passability index relationship model characterizes the quantitative mapping relationship between the thermal inertia of the fire soil and at least one mechanical parameter used to evaluate the passability of the support.

[0111] Specifically, this part of the model is pre-constructed on the Earth's surface. The specific construction process is as follows: multiple sets of simulated fire-soil samples are prepared, and the measured values ​​of thermal inertia and mechanical parameters of the simulated fire-soil samples are obtained; a dataset is constructed based on the measured values ​​of thermal inertia and mechanical parameters, and the dataset is divided into a training set, a validation set, and a test set; a multinomial regression model is determined, and the multinomial regression model is trained based on the training set to obtain an initial model; or a support vector regression model or a gradient boosting regression tree model is used, and the support vector regression model or the gradient boosting regression tree model is trained based on the training set to obtain an initial model; the initial model is cross-validated based on the validation set, and the model is optimized based on the test set to obtain a model showing the relationship between thermal inertia and passability index.

[0112] For evaluating the mechanical parameters of support passability, this invention preferentially uses the conic index. CI The model relating thermal inertia to throughput index is specifically a quantitative model relating thermal inertia to the conic index, i.e. CI =f( I i ).

[0113] The relationship model between thermal inertia and throughput index is established in advance through the following steps:

[0114] 1. Ground simulation test: Prepare in a ground test field simulating the Martian environment. N A group of simulated fire soil samples with different physical properties, and N ≥3; the physical properties include one or more of particle size, density, and component ratio; wherein, the simulated fire soil sample can be: JLU Mars series fire soil, specifically JLU Mars1 (small particle size fire soil), JLU Mars2 (medium particle size fire soil), and JLU Mars3 (large particle size fire soil). The density has at least four different densities, and then a relationship model between simulated fire soil of different particle sizes or densities, thermal inertia, and conic index is established, because the rover's passability on fire soil of different particle sizes and densities is different, so this model is established in advance.

[0115] 2. Synchronous Parameter Measurement: For each group of simulated fire-soil samples, under the same environmental conditions, the measured values ​​of thermal inertia and mechanical parameters are measured synchronously; the measured mechanical parameters include at least the conic index. CI In ground environments, use the formula .

[0116] In a simulated Martian environment chamber (which realistically simulates the Martian environment, specifically: temperature -125℃ to 20℃, atmospheric pressure 610 Pa, CO2 content ≥95%), thermocouples were inserted into simulated soil to measure the temperature response curve. A constant heating power was applied to the simulated Martian soil for 20-30 minutes, and the temperature change curve over time was recorded. Then, a heat conduction model was used to fit the temperature curve, and the thermal conductivity of the material was calculated using the fitted parameters. .density The mass and volume of the sample, and its specific heat capacity, are measured using the weighing method. The measurements were taken directly using a differential scanning calorimeter.

[0117] 3. Model training: using the measured values ​​of thermal inertia of all samples. I i As input, the corresponding measured values ​​of mechanical parameters (such as...) CI The output is ), and the quantitative mapping relationship model f is obtained by training with a data fitting algorithm or a machine learning algorithm.

[0118] The thermal inertia-passability index relationship model CI =f( I i The establishment of the model is achieved through the following steps. First, it is necessary to ensure that the model has good fitting accuracy and generalization ability:

[0119] 1. Dataset Preparation: After completing the ground simulation experiment and acquiring... Measured values ​​of thermal inertia of simulated volcanic soil samples I i With conic index CI i Then, the data is organized into a matrix form:

[0120] ;

[0121] in, It is the set of measured values ​​of thermal inertia. For the first Measured value of thermal inertia For matrix transpose, It is the set of conical indices. For the first A conic index.

[0122] The dataset was then divided into a training set Dtrain (70%), a validation set Dval (15%), and a test set Dtest (15%).

[0123] 2. Data preprocessing: processing the input features I i With output labels CI Standardize them separately, and the expressions are as follows:

[0124] ;

[0125] in, For standardization I , for I The mean, The initial standard deviation, For standardization CI , for CI The mean, for CI The standard deviation is used to improve the numerical stability of model training.

[0126] 3. Model selection and training: This invention provides two optional model strategies:

[0127] a. Based on traditional regression fitting methods, a multinomial regression model is selected, the general form of which is:

[0128] ;

[0129] in: K Let the number of stages be the polynomial number. For the first The coefficients of the term's thermal inertia are solved using the least squares method. And use the validation set to determine the optimal number of stages K. opt .

[0130] b. Machine learning-based methods: Training is performed using Support Vector Regression (SVR) or Gradient Boosting Regression Tree (GBRT):

[0131] For SVR training: use radial basis functions. As a kernel function, the regularization parameter C and kernel coefficient γ are optimized on the validation set through network search;

[0132] For GBRT training: Set the maximum tree depth, learning rate, and number of trees, use gradient descent for iterative optimization, and stop early on the validation set to prevent overfitting.

[0133] Furthermore, for model validation and performance evaluation: cross-validation was performed using the validation set Dval, and the evaluation metrics included:

[0134] Root mean square error: ;

[0135] in, For the sample size, For the first j The true value of the conic index for each sample. For the first j The predicted value of the conic index for each sample.

[0136] Coefficient of determination: ;

[0137] Then, based on the above indicators, the optimal model structure and parameter combination on the validation set are selected.

[0138] Model Testing and Consolidation: The generalization performance of the final model is evaluated using an independent test set, Dtest. If the test results meet the pre-accuracy requirements, the model parameters are consolidated into the non-volatile memory of the Mars rover's data processing module, forming a callable mapping function f, thus obtaining a model relating thermal inertia and passability indicators.

[0139] Then, the estimated value of the thermal inertia of the fire soil is input into the relationship model between thermal inertia and passability index, and the predicted passability index is output.

[0140] Step S40: Determine the passability status level based on the predicted passability index, and generate navigation control suggestions based on the passability status level, so that the Mars rover can pass through the target area according to the Mars rover navigation control suggestions.

[0141] For accessibility assessment and decision-making, this invention compares the predicted accessibility index with a preset safety threshold, determines the accessibility status of the target area based on the comparison result, and generates corresponding navigation control suggestions.

[0142] Specifically, a preset safety margin and a preset safety threshold are determined; if the predicted passability index is greater than or equal to the sum of the preset safety margin and the preset safety threshold, the target area is determined to be at a first passability status level; if the predicted passability index is greater than the preset safety threshold but less than the sum of the preset safety margin and the preset safety threshold, the target area is determined to be at a second passability status level; if the predicted passability index is less than the preset safety threshold, the target area is determined to be at a third passability status level; corresponding navigation control suggestions are generated based on the first passability status level, the second passability status level, and the third passability status level, respectively, so that the Mars rover can pass through the target area according to the Mars rover navigation control suggestions.

[0143] It is understood that the passability judgment in this invention specifically refers to: if the predicted conic index... CI Greater than or equal to the minimum permissible conic exponent threshold of the Mars rover CI min (wherein, the minimum allowable conic exponent threshold) CI The value of min is determined by the rover's wheel parameters, mass distribution, and safety factor, with a typical value of 5~10 N / m² (preferably set to 6.0 N / m² in this invention). If the value is less than 5, the target area is considered passable; otherwise, it is considered a high-risk area and avoidance is recommended.

[0144] Regarding the Mars rover's navigation and control recommendations, the specific process is as follows:

[0145] 1. Passability Classification: Based on the comparison between the predicted conic index and the preset safety threshold, the passability of the target area is divided into three levels: Level 1 (Safe to Pass), the predicted conic index is greater than or equal to the sum of the safety threshold and the safety margin; Level 2 (Critical Risk), the predicted conic index is between the safety threshold and the safety margin; Level 3 (High Risk, Impossible), the predicted conic index is less than the safety threshold. The safety margin is determined based on the Mars rover's design parameters and ground calibration tests, and its value ranges from 1 to 2 N / m. 2 .

[0146] 2. Generate corresponding navigation control suggestions based on different levels of passability:

[0147] Level 1 (Safe to Pass) Area: Recommended Action: Normal passage; Driving Parameters: Maintain the currently set speed and direction; Monitoring Mode: Use the normal detection frequency.

[0148] Level 2 (Critical Risk) Zone: Recommended Actions: Proceed cautiously or tentatively; Driving Parameters: Reduce vehicle speed to 30% to 50% of the original set speed and maintain straight driving, avoiding emergency steering or braking; Monitoring Mode: Enable enhanced monitoring to collect wheel slip rate, drive motor torque, and wheel sinking depth in real time; Safe Retreat Strategy: If the slip rate exceeds the preset upper limit or the sinking depth exceeds the allowable value, an automatic retreat action will be triggered.

[0149] Level 3 (High-risk, impassable) area: Recommended actions: Avoid; Path replanning: Using this area as an obstacle, replan the detour route based on the rover's current location and local environment map using a path search algorithm; Map update: Mark this area as a permanent obstacle and update the onboard navigation map.

[0150] 3. Multi-factor decision fusion: To improve the adaptability and reliability of the recommendations, the system integrates the following information when generating recommendations: a) Terrain slope information: When the slope of the target area exceeds the set threshold, its passability status level is downgraded by one level; b) Scientific mission priority: If the target area is identified as a high-value scientific exploration target, recommendations containing extremely low-speed trial passage instructions are generated; c) System energy status: When the remaining power is lower than the set threshold, recommendations tending towards the shortest safe path are generated first to reduce energy consumption; d) Historical driving data: Based on the rover's past records of passing through similar terrains, the safety margin or status classification threshold is adaptively adjusted.

[0151] Beneficial effects:

[0152] 1. This invention enables the rover to assess the passability of unknown terrain before physical contact with the Mars rover;

[0153] 2. This invention indirectly perceives the deep mechanical properties of soil (such as density and cohesion) that are directly related to support and passage through the integral characteristic of thermal inertia. It overcomes the limitation of visual methods that can only perceive surface geometry and has a unique ability to identify visually deceptive terrains such as thin shells and loose soil.

[0154] 3. The core sensor in this invention is the thermal infrared camera commonly used on Mars rovers, eliminating the need for additional large and complex dedicated mechanical detection equipment, which meets the stringent requirements of deep space exploration missions for payload weight and reliability.

[0155] 4. The Martian atmosphere is thin, dry, and devoid of vegetation, resulting in a relatively simple surface heat exchange process. This actually helps to improve the accuracy of thermal inertia retrieval based on temperature changes, making this invention more advantageous for application in the Martian environment than on Earth.

[0156] In addition, this invention optimizes the passive monitoring window by studying time-controlled factors and innovatively proposes an active thermal induction mode, which can quickly excite and monitor the thermal response of the soil under mission time constraints, achieving rapid assessment and meeting the needs of real-time navigation of the Mars rover.

[0157] Furthermore, such as Figure 2 As shown, based on the above-mentioned non-contact Mars rover passability prediction method based on thermal inertia, the present invention also provides a non-contact Mars rover passability prediction system based on thermal inertia, wherein the non-contact Mars rover passability prediction system based on thermal inertia includes:

[0158] The regional data acquisition module 51 is used to determine the target area and acquire the surface temperature time series and environmental parameters of the target area;

[0159] The thermal inertia estimation module 52 is used to construct a thermal inertia estimation model and input the surface temperature time series and the environmental parameters into the thermal inertia estimation model to obtain the thermal inertia estimation value of the fire soil;

[0160] The passability index output module 53 is used to construct a relationship model between thermal inertia and passability index, input the estimated value of the fire soil thermal inertia into the relationship model between thermal inertia and passability index, and output the predicted passability index.

[0161] The navigation suggestion generation module 54 is used to determine the passability status level based on the predicted passability index, and generate navigation control suggestions based on the passability status level, so that the rover can pass through the target area according to the rover navigation control suggestions.

[0162] Furthermore, such as Figure 3 As shown, based on the above-mentioned non-contact Mars rover passability prediction method and system based on thermal inertia, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 3 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0163] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the terminal. Further, the memory 20 may include both internal and external storage units of the terminal. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a Mars rover non-contact passability prediction program 40 based on thermal inertia, which can be executed by the processor 10 to implement the Mars rover non-contact passability prediction method based on thermal inertia in this application.

[0164] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the non-contact passability prediction method for Mars rovers based on thermal inertia.

[0165] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface.

[0166] In one embodiment, when the processor 10 executes the thermal inertia-based rover non-contact passability prediction program 40 in the memory 20, it implements the steps of the thermal inertia-based rover non-contact passability prediction method as described above.

[0167] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a Mars rover non-contact passability prediction program based on thermal inertia, and the Mars rover non-contact passability prediction program based on thermal inertia, when executed by a processor, implements the steps of the Mars rover non-contact passability prediction method based on thermal inertia as described above.

[0168] In summary, this invention provides a non-contact method, system, terminal, and storage medium for predicting the passability of a Mars rover based on thermal inertia. The method includes: determining a target area and acquiring the surface temperature time series and environmental parameters of the target area; constructing a thermal inertia estimation model and inputting the surface temperature time series and environmental parameters into the thermal inertia estimation model to obtain an estimated value of the soil thermal inertia; constructing a relationship model between thermal inertia and passability indicators, inputting the estimated value of the soil thermal inertia into the thermal inertia-passability indicator relationship model, and outputting a predicted passability indicator; determining the passability status level based on the predicted passability indicator, and generating navigation control suggestions based on the passability status level, enabling the Mars rover to pass through the target area according to the Mars rover navigation control suggestions. This invention, by constructing a thermal inertia estimation model, can extract the estimated value of the thermal inertia of the target area and predict the passability indicator of the Mars rover based on the estimated value of the thermal inertia, which can effectively improve the assessment accuracy of the Mars rover passing through the target area and ensure the driving safety of the Mars rover.

[0169] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0170] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0171] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A non-contact method for predicting the passability of a Mars rover based on thermal inertia, characterized in that, The non-contact mobility prediction method for Mars rovers based on thermal inertia includes: Identify the target area and obtain the time series of surface temperature and environmental parameters of the target area; The process of determining the target area and obtaining the land surface temperature time series and environmental parameters of the target area specifically includes: Determine the target region, obtain the region shape of the target region, and determine the time resolution based on the region shape; A preset illumination mode is determined, and the thermal infrared imaging equipment on the Mars rover is used to detect thermal radiation in the target area according to the time resolution and the preset illumination mode to obtain multiple frames of thermal infrared images of the target area within a preset time period. The preset lighting modes include natural lighting mode and active thermal radiation mode; Extract the time sequence of surface temperature variation from the multi-frame thermal infrared images to obtain the surface temperature time series; Obtain the environmental parameters of the target area; A thermal inertia estimation model was constructed, and the surface temperature time series and the environmental parameters were input into the thermal inertia estimation model to obtain the estimated value of the thermal inertia of the fire-soil mixture. A model relating thermal inertia to passability index is constructed. The estimated thermal inertia of the fire soil is input into the model relating thermal inertia to passability index, and the predicted passability index is output. The passability status level is determined based on the predicted passability index, and navigation control suggestions are generated based on the passability status level, so that the rover can pass through the target area according to the rover navigation control suggestions.

2. The non-contact Mars rover passability prediction method based on thermal inertia according to claim 1, characterized in that, The step of extracting the time sequence of surface temperature changes from the multi-frame thermal infrared images to obtain the surface temperature time series specifically includes: The multi-frame thermal infrared images are preprocessed, including fixed defect detection processing, noise reduction processing, alignment processing, and feature point extraction processing, to obtain a preprocessed image. Calculate the mean and standard deviation of the temperature of the effective pixels in each frame of the preprocessed image, and obtain the target temperature value based on the mean and standard deviation of the temperature; The target temperature values ​​are arranged in chronological order to obtain the initial temperature time series; The initial temperature time series is smoothed to obtain the surface temperature time series.

3. The non-contact Mars rover passability prediction method based on thermal inertia according to claim 1, characterized in that, The environmental parameters include albedo; The acquisition of environmental parameters of the target area specifically includes: The target area is imaged and radiometrically calibrated using a multispectral camera on the Mars rover to obtain an albedo image. The average albedo in the albedo image is then calculated to obtain the albedo. Alternatively, a high-resolution visible light camera or color camera carried by a Mars rover can be used to image and perform radiometric calibration on the terrain areas with shadow boundaries in the target area to obtain a reflectance image; The reflectance image is divided into a direct illumination area and a shadow area using an image processing algorithm, and the average reflectance of the direct illumination area and the shadow area is calculated. The average reflectance of the shaded area is corrected for atmospheric scattering to obtain the albedo. Alternatively, a pre-trained deep convolutional neural network can be constructed to obtain a visible light image of the target region, and the visible light image can be analyzed using the pre-trained deep convolutional neural network to obtain the albedo.

4. The non-contact Mars rover passability prediction method based on thermal inertia according to claim 3, characterized in that, The construction of the thermal inertia estimation model, and the input of the surface temperature time series and the environmental parameters into the thermal inertia estimation model to obtain the estimated value of the volcanic thermal inertia, specifically includes: A thermal inertia estimation model is constructed based on the Martian surface heat balance equation and heat conduction equation. The surface temperature time series and the environmental parameters are input into the thermal inertia estimation model, and the soil thermal inertia is calculated based on the surface temperature time series and the environmental parameters through the thermal inertia estimation model to obtain the estimated value of Martian soil thermal inertia. The expression for the thermal inertia estimation model is as follows: ； in, This is an estimated value for the thermal inertia of volcanic soil. This is a comprehensive correction factor related to the environment of the target area. The albedo is one of the environmental parameters. The maximum surface temperature difference within a selected time period in the surface temperature time series.

5. The non-contact Mars rover passability prediction method based on thermal inertia according to claim 1, characterized in that, The construction of the relationship model between thermal inertia and passability index involves inputting the estimated thermal inertia of the fire-soil mixture into the model and outputting a predicted passability index, specifically including: Multiple sets of simulated fire soil samples were prepared, and the measured values ​​of thermal inertia and mechanical parameters of the simulated fire soil samples were obtained. A dataset is constructed based on the measured values ​​of thermal inertia and mechanical parameters, and the dataset is divided into a training set, a validation set, and a test set. Determine a multinomial regression model, and train the multinomial regression model using the training set to obtain an initial model; Alternatively, a support vector regression model or a gradient boosting regression tree model can be used. The support vector regression model or the gradient boosting regression tree model can be trained based on the training set to obtain an initial model. The initial model is cross-validated based on the validation set, and the model is optimized based on the test set to obtain a model relating thermal inertia and passability index. The estimated thermal inertia of the fire and soil is input into the relationship model between thermal inertia and passability index, and the predicted passability index is output.

6. The non-contact Mars rover passability prediction method based on thermal inertia according to claim 1, characterized in that, The step of determining the passability status level based on the predicted passability index and generating navigation control suggestions based on the passability status level, enabling the Mars rover to pass through the target area according to the Mars rover navigation control suggestions, specifically includes: Determine the preset safety margin and preset safety threshold; If the predicted passability index is greater than or equal to the sum of the preset safety margin and the preset safety threshold, then the target area is determined to be at the first passability status level. If the predicted passability index is greater than the preset safety threshold and less than the sum of the preset safety margin and the preset safety threshold, then the target area is determined to be at the second passability status level. If the predicted passability index is less than the preset safety threshold, the target area is determined to be at the third passability status level. Based on the first passability status level, the second passability status level, and the third passability status level, corresponding navigation control suggestions are generated respectively, enabling the Mars rover to pass through the target area according to the Mars rover navigation control suggestions.

7. A non-contact passability prediction system for Mars rovers based on thermal inertia, characterized in that, The thermal inertia-based non-contact rover passability prediction system is used to implement the thermal inertia-based non-contact rover passability prediction method as described in any one of claims 1-6, wherein the thermal inertia-based non-contact rover passability prediction system comprises: The regional data acquisition module is used to determine the target area and acquire the land surface temperature time series and environmental parameters of the target area; The thermal inertia estimation module is used to construct a thermal inertia estimation model and input the surface temperature time series and the environmental parameters into the thermal inertia estimation model to obtain the thermal inertia estimation value of the fire soil; The passability index output module is used to construct a model relating thermal inertia and passability index. The estimated value of the fire soil thermal inertia is input into the model relating thermal inertia and passability index, and the predicted passability index is output. The navigation suggestion generation module is used to determine the passability status level based on the predicted passability index, and generate navigation control suggestions based on the passability status level, so that the rover can pass through the target area according to the rover navigation control suggestions.

8. A terminal, characterized in that, The terminal includes: a memory, a processor, and a thermal inertia-based non-contact Mars rover passability prediction program stored in the memory and executable on the processor. When the thermal inertia-based non-contact Mars rover passability prediction program is executed by the processor, it implements the steps of the thermal inertia-based non-contact Mars rover passability prediction method as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a Mars rover non-contact passability prediction program based on thermal inertia, which, when executed by a processor, implements the steps of the Mars rover non-contact passability prediction method based on thermal inertia as described in any one of claims 1-6.

Images